As businesses increasingly integrate AI language models into their operations, cybercriminals are evolving to monetize these tools. Flare has discovered that over 200,000 credentials to AI language models have surfaced on the dark web, heightening the risk of sensitive data leaks.
During the webinar, experts Jason Haddix (BuddoBot CISO), Serge-Olivier Paquette (Flare Director of Innovation), and Eric Clay (Flare VP of Marketing) shed light on:
- The use of open-source models, such as LLAMA and Vicuna, where public weights allow malicious actors to bypass safety mechanisms.
Stay informed and prepared by understanding the evolving AI threats in the cyber landscape.